Improve performance of slice_strings for long strings

[davidwendt]

Description

Improves on performance for longer strings with cudf::strings::slice_strings() API. The cudf::string_view::substr was reworked to minimize counting characters and the gather version of make_strings_children is used to build the resulting strings column. This version is already optimized for small and large strings.

Additionally, the code was refactored so the common case of step==1 and start < stop can also make use of the gather approach. Common code was also grouped closer together to help navigate the source file better.

The slice.cpp benchmark was updated to better measure large strings with comparable slice boundaries. The benchmark showed performance improvement was up to 9x for larger strings with no significant degradation for smaller strings.

Reference #13048 and #12445

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[davidwendt]

From slice benchmarks

rows	width	baseline	new	x
4096	32	0.117	0.132	0.9
4096	64	0.142	0.137	1.0
4096	128	0.199	0.148	1.3
4096	256	0.330	0.176	1.9
4096	512	0.603	0.233	2.6
4096	1024	1.26	0.319	3.9
4096	2048	2.64	0.553	4.8
4096	4096	5.55	1.02	5.4
4096	8192	12.4	2.54	4.9
32768	32	0.119	0.135	0.9
32768	64	0.147	0.141	1.0
32768	128	0.210	0.156	1.3
32768	256	0.363	0.189	1.9
32768	512	0.669	0.271	2.5
32768	1024	1.50	0.390	3.8
32768	2048	3.56	0.850	4.2
32768	4096	7.46	1.63	4.6
32768	8192	14.8	3.14	4.7
262144	32	0.190	0.171	1.1
262144	64	0.324	0.199	1.6
262144	128	1.38	0.288	4.8
262144	256	7.38	1.47	5.0
262144	512	22.7	5.18	4.4
262144	1024	52.9	5.80	9.1
262144	2048	139	18.9	7.4
262144	4096	293	39.2	7.5
2097152	32	0.780	0.479	1.6
2097152	64	1.72	0.597	2.9
2097152	128	8.54	1.10	7.8
2097152	256	53.2	7.34	7.2
2097152	512	160	32.4	4.9
16777216	32	5.75	2.82	2.0
16777216	64	13.1	3.74	3.5

[davidwendt]

/merge